摘要
建立了坝体沉降变形分解的统一模型,采用最小二乘法进行多元线性回归分析,将面板堆石坝坝体沉降变形的实测值分解为荷载作用变形与蠕变作用变形。运用MATLAB语言提供的神经网络功能函数,建立面板堆石坝反分析的BP神经网络模型,对邓肯E-B模型的材料参数进行反演计算。实例计算结果表明,该方法简单、有效、可靠,计算结果满足工程精度要求,可应用于实际工程。
The united model of settlement deformation decomposition of dam hay been built, the polyatomic linear regression analyses are carried out by using the least square method, the observed total settlement deformations of concrete face rockfill dams （CFRD） are decomposed into ones of loads effect and creep effect. BP ANN model of back analysis of CFRD has been constructed with Newff Function of MATLAB language, the material parameters of Duncan E-B Model is calculated by back analysis. The result shows that neural networks method of back analysis is simple, effective and reliable, and satisfies the demand of precision, can be applied to projects.
出处
《中国农村水利水电》
北大核心
2008年第7期64-67,共4页
China Rural Water and Hydropower
关键词
面板堆石坝
反分析
BP神经网络方法
邓肯E-B模型
变形分解
concrete face roekfill darm inverse analysis
BP neural networks method
Duncan E-B Model
deformation decomposition
